Brush gear mechanisms are used extensively in certain motors, generators and other rotating electrical machinery to conduct current between stationary and rotating parts (commutators or slip rings). To avoid or at best minimise the wear of the rotating parts, which are typically constructed of copper or brass, brushes are made of soft materials, such as carbon and graphite, so as to preferentially wear. It is important that brush wear is monitored to ensure that the brushes do not wear beyond a minimum level as a completely worn brush can lead to electrical arcing resulting in machine part failure.
One established practice of monitoring brush wear is to install brush assemblies in brush houses with visual inspection means to enable operator inspection. Due to human factors however machine failures still occur. For other reasons, including cost reduction it maybe desirable to at least supplement such monitoring with automated monitoring. As a result instrumented brush-monitoring systems have been developed.
Monitoring of the excitation voltage supplied a brush is one known method of monitoring brush wear. For example GB1564384 describes a process of monitoring the signature of the excitation voltage for unusual high-energy noise spikes created by brush arcing. It does this by using sensors connected to each brush and employing comparative algorithms and filtering means to analysis the data. U.S. Pat. No. 4,451,786 and CA 1194112 also teach of similar methods with similar disadvantages including that a fault condition i.e. arcing but not brush wear, is monitored and so the method is reactive rather than preventative. Further disadvantages include the fact that monitoring involves complex signature filtering and interpretation and so is open to error. The methods also inherently lack the ability to utilize one sensor to monitor multiple brushes.
JP6141513 discloses another brush monitoring method that utilizes brush electrical signatures relying only on inputs of armature current and an armature revolution count to calculate brush wear. As the method does not directly measure or detect brush wear the accuracy of the method is limited and yet despite analyses simplicity still requires a sensor per monitored brush.
GB 1 215 649 A discloses a further brush monitoring system having a feeler portion, supported in a frame mounted on a brush holder, resiliently urged onto the brush. When the outer end of the brush passes beyond a predetermined portion, due to brush wear, the feeler portion actuates a switch in a monitoring circuit for that brush.
Larger machines typically have 30 or more brushes and therefore to improve inherent monitoring reliability and simplify maintenance and installation it is advantageous for these systems to have as few sensors as possible. The requirement for the cited references teaches to have a sensor per monitored brush therefore disadvantages all the above teachings.
DE 3417711 A1 provides another brush wear monitoring system capable of monitoring the wear of multiple brushes. The switching arrangement of the system is mounted on each brush holder and comprises a swivelled lever that is biased onto the brush. The lever further comprises a contact that when the brush wears, moves towards a brush holder, holding the brush, by the swivel motion of the level. When the brush wears to a certain point the contact contacts the brush holder. This has the effect of closing the brush wear recording circuit so by enabling the triggering of an brush wear alarm.
DE 86 00 934 U1 provides another brush wear monitoring system capable of monitoring the wear of multiple brushes by using a swivel arm that swivels with the wear movement of the brush so as to close a switch at a brush wear point.
DE 197 58 235 A1 provides another brush wear monitoring system comprising of a rod mounted on a brush. The system uses the principle of a sensor detecting a source mounted on the rod as the rod is moved by brush wear. The source may be a magnetised ring while the sensor may be either a magnetic-hydraulic sensor or an optical-hydraulic sensor.